Unlocking echocardiogram measurements for heart disease research through natural language processing
Abstract Background In order to investigate the mechanisms of cardiovascular disease in HIV infected and uninfected patients, an analysis of echocardiogram reports is required for a large longitudinal multi-center study. Implementation A natural language processing system using a dictionary lookup,...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
BMC
2017-06-01
|
Series: | BMC Cardiovascular Disorders |
Subjects: | |
Online Access: | http://link.springer.com/article/10.1186/s12872-017-0580-8 |
_version_ | 1811206487581655040 |
---|---|
author | Olga V. Patterson Matthew S. Freiberg Melissa Skanderson Samah J. Fodeh Cynthia A. Brandt Scott L. DuVall |
author_facet | Olga V. Patterson Matthew S. Freiberg Melissa Skanderson Samah J. Fodeh Cynthia A. Brandt Scott L. DuVall |
author_sort | Olga V. Patterson |
collection | DOAJ |
description | Abstract Background In order to investigate the mechanisms of cardiovascular disease in HIV infected and uninfected patients, an analysis of echocardiogram reports is required for a large longitudinal multi-center study. Implementation A natural language processing system using a dictionary lookup, rules, and patterns was developed to extract heart function measurements that are typically recorded in echocardiogram reports as measurement-value pairs. Curated semantic bootstrapping was used to create a custom dictionary that extends existing terminologies based on terms that actually appear in the medical record. A novel disambiguation method based on semantic constraints was created to identify and discard erroneous alternative definitions of the measurement terms. The system was built utilizing a scalable framework, making it available for processing large datasets. Results The system was developed for and validated on notes from three sources: general clinic notes, echocardiogram reports, and radiology reports. The system achieved F-scores of 0.872, 0.844, and 0.877 with precision of 0.936, 0.982, and 0.969 for each dataset respectively averaged across all extracted values. Left ventricular ejection fraction (LVEF) is the most frequently extracted measurement. The precision of extraction of the LVEF measure ranged from 0.968 to 1.0 across different document types. Conclusions This system illustrates the feasibility and effectiveness of a large-scale information extraction on clinical data. New clinical questions can be addressed in the domain of heart failure using retrospective clinical data analysis because key heart function measurements can be successfully extracted using natural language processing. |
first_indexed | 2024-04-12T03:48:20Z |
format | Article |
id | doaj.art-25e6fb3ae444449d8fe4e55d9510420a |
institution | Directory Open Access Journal |
issn | 1471-2261 |
language | English |
last_indexed | 2024-04-12T03:48:20Z |
publishDate | 2017-06-01 |
publisher | BMC |
record_format | Article |
series | BMC Cardiovascular Disorders |
spelling | doaj.art-25e6fb3ae444449d8fe4e55d9510420a2022-12-22T03:49:03ZengBMCBMC Cardiovascular Disorders1471-22612017-06-0117111110.1186/s12872-017-0580-8Unlocking echocardiogram measurements for heart disease research through natural language processingOlga V. Patterson0Matthew S. Freiberg1Melissa Skanderson2Samah J. Fodeh3Cynthia A. Brandt4Scott L. DuVall5Department of Veterans Affairs Salt Lake City Health Care SystemVA Tennessee Valley Health Care SystemConnecticut VA Healthcare SystemCenter for Medical Informatics, School of Medicine, Yale UniversityConnecticut VA Healthcare SystemDepartment of Veterans Affairs Salt Lake City Health Care SystemAbstract Background In order to investigate the mechanisms of cardiovascular disease in HIV infected and uninfected patients, an analysis of echocardiogram reports is required for a large longitudinal multi-center study. Implementation A natural language processing system using a dictionary lookup, rules, and patterns was developed to extract heart function measurements that are typically recorded in echocardiogram reports as measurement-value pairs. Curated semantic bootstrapping was used to create a custom dictionary that extends existing terminologies based on terms that actually appear in the medical record. A novel disambiguation method based on semantic constraints was created to identify and discard erroneous alternative definitions of the measurement terms. The system was built utilizing a scalable framework, making it available for processing large datasets. Results The system was developed for and validated on notes from three sources: general clinic notes, echocardiogram reports, and radiology reports. The system achieved F-scores of 0.872, 0.844, and 0.877 with precision of 0.936, 0.982, and 0.969 for each dataset respectively averaged across all extracted values. Left ventricular ejection fraction (LVEF) is the most frequently extracted measurement. The precision of extraction of the LVEF measure ranged from 0.968 to 1.0 across different document types. Conclusions This system illustrates the feasibility and effectiveness of a large-scale information extraction on clinical data. New clinical questions can be addressed in the domain of heart failure using retrospective clinical data analysis because key heart function measurements can be successfully extracted using natural language processing.http://link.springer.com/article/10.1186/s12872-017-0580-8Natural language processingText miningInformation extractionEchocardiographyHeart functionLeft ventricular ejection fraction |
spellingShingle | Olga V. Patterson Matthew S. Freiberg Melissa Skanderson Samah J. Fodeh Cynthia A. Brandt Scott L. DuVall Unlocking echocardiogram measurements for heart disease research through natural language processing BMC Cardiovascular Disorders Natural language processing Text mining Information extraction Echocardiography Heart function Left ventricular ejection fraction |
title | Unlocking echocardiogram measurements for heart disease research through natural language processing |
title_full | Unlocking echocardiogram measurements for heart disease research through natural language processing |
title_fullStr | Unlocking echocardiogram measurements for heart disease research through natural language processing |
title_full_unstemmed | Unlocking echocardiogram measurements for heart disease research through natural language processing |
title_short | Unlocking echocardiogram measurements for heart disease research through natural language processing |
title_sort | unlocking echocardiogram measurements for heart disease research through natural language processing |
topic | Natural language processing Text mining Information extraction Echocardiography Heart function Left ventricular ejection fraction |
url | http://link.springer.com/article/10.1186/s12872-017-0580-8 |
work_keys_str_mv | AT olgavpatterson unlockingechocardiogrammeasurementsforheartdiseaseresearchthroughnaturallanguageprocessing AT matthewsfreiberg unlockingechocardiogrammeasurementsforheartdiseaseresearchthroughnaturallanguageprocessing AT melissaskanderson unlockingechocardiogrammeasurementsforheartdiseaseresearchthroughnaturallanguageprocessing AT samahjfodeh unlockingechocardiogrammeasurementsforheartdiseaseresearchthroughnaturallanguageprocessing AT cynthiaabrandt unlockingechocardiogrammeasurementsforheartdiseaseresearchthroughnaturallanguageprocessing AT scottlduvall unlockingechocardiogrammeasurementsforheartdiseaseresearchthroughnaturallanguageprocessing |